Project Summary/Abstract Most prostate cancer deaths are due to failed treatment response or to metastasis, particularly to bone. We have been studying the mechanisms responsible for treatment failure using genetically-engineered mouse models (GEMMs) that recapitulate key features of advanced prostate cancer, including castration-resistant prostate cancer (CRPC) and aggressive-variant CRPC with neuroendocrine differentiation (CRPC-NE), and that develop highly penetrant metastatic prostate cancer including to bone, which is the primary site of metastasis in humans. Using novel cross-species computational approaches, we have shown that the mechanisms underlying disease progression in these GEMMs are conserved with human prostate cancer, while drug response in GEMMs can be predictive of drug response in humans. Among our major findings, we have shown that GEMMs based on combined loss-of-function of Pten and p53 (NPp53), which are frequently co-mutated in human CRPC, model key phenotypic and molecular features of treatment-induced aggressive variant CRPC. Not only do these NPp53 mice fail to respond to treatment with anti-androgens, treatment actually accelerates disease progression, which we have called exceptional non-responders. Furthermore, we have shown that the treatment-induced neuroendocrine phenotype (CRPC-NE) of these exceptional non-responders is related to lineage plasticity. Thus, our proposed studies will test the hypotheses that: (i) co-clinical analyses of GEMMs and human prostate cancer can elucidate biological and molecular mechanisms of drug response, and (ii) lineage plasticity is an important novel mechanism of drug resistance. In Aim 1 we will perform co-clinical investigations capitalizing on our GEMMs that model key aspects of advanced prostate cancer, and complemented with analyses of human prostate cancer organoid models. We will evaluate the efficacy of clinically-relevant drugs and drug combinations, focusing on those that: (a) counteract cellular plasticity associated with treatment resistance in CRPC; and (b) target bone metastasis. In Aim 2, we will investigate molecular mechanisms of drug response leveraging our genome-wide regulatory networks that enable cross-species integration between data from GEMMs and human prostate cancer. We will focus on identifying ?treatment response regulators? that inform on response to: (i) lineage plasticity and/or (ii) bone metastasis. In Aim 3 we will identify novel drivers of advanced prostate cancer using a forward genetic screening approach. Toward this end, we have undertaken a genetic screen utilizing the Sleeping Beauty (SB) murine transposon-based system, and have shown that mice harboring the activated transposon display accelerated lethal prostate cancer phenotypes. Taken together, the successful implementation of these Aims will identify tumor contexts that are responsive to drug treatment and molecular mechanisms of drug response, and will identify new targets for intervention.